Gabb

16 posts

Gabb banner
Gabb

Gabb

@gabb0347

Norwegian engineer and tinkerer living in the US. 🇳🇴🇺🇸🌍

San Francisco, CA Katılım Haziran 2026
42 Takip Edilen54 Takipçiler
Gabb
Gabb@gabb0347·
Gabb tweet media
ZXX
0
0
0
38
Humanoid Scott
Humanoid Scott@GoingBallistic5·
The maxim, “Prototypes are easy, production is hard”, does not apply to humanoids No one, including Tesla, has a working prototype worth scaling in the 10’s of 1000’s, let alone millions It’s not a scaling problem, is a prototype problem Humanoid prototypes are hard Current prototypes might be scalable but are not sufficiently capable nor robust. Nor safe. Including Optimus You don’t scale a crappy humanoid. Unitree could easily scale the G1 to millions, but to what ends? Obsolescence in 6 months? Humanoid prototypes are hard Ramping up a million and 10 million annual capacity lines simultaneously is premature. We will never see that quantity of V3’s, not because the supply chain isn’t yet mature, but because the prototype isn’t Tesla is attempting to scale a moving target Tesla is impressing by flexing its scaling muscles first, rather than demonstrating any clear technological adroitness first V3 will not be revealed this year, not over piracy concerns, but because the bar for humanoid demos is beyond its current capabilities to demonstrate The million plus Optimi will be V5 or higher Humanoid prototypes worth scaling are hard
Humanoid Scott tweet media
Dillon Loomis@DillonLoomis

Who thinks Tesla is behind with Optimus? I get the "we don't want our competitors to copy us" argument for demos, but couldn't Tesla just put some gloves and a shirt over Optimus and give us a quick update into what they have cooking? It's been QUIET out there from the Optimii Fwiw I don't think Tesla's behind, I think part of it is optimizing differently and taking different paths to market. NEO is going teleoperator route to ship to customer homes for training. Tesla will likely do plenty of learning in the factories with initial deployment there. The list of differences goes on tbh But the great news is 1X is an American-Norwegian company (main operations in America, founded in Norway) and the faster the market has options from American manufacturers, the better (some motors come from Norway). The main 1X factory currently is in Hayward, CA (~58k sq. ft. + 10k unit capacity/yr.) but a bigger location is set to come online in San Carlos later this year (~231k sq. ft. to hit 100k units/yr.) Plus the VP of Operations is a former SpaceX employee (Vikram Kothari - he spent 8 years there and nearly a decade at Microsoft in hw supply chain leadership)

English
82
25
336
92.4K
Gabb
Gabb@gabb0347·
@waqasmakes @radbackwards Consistency compounds and builds momentum. Momentum is what real mastery is made of.
English
1
0
1
43
dar
dar@radbackwards·
Obsession is the only path to greatness
English
66
218
1.7K
68.4K
clankr
clankr@clankrmedia·
Robot hands got serious. Here are four dexterous hands: 🔵 Tesla Optimus V3: 22 DOF, motors in the forearm. 🔵 BrainCo Revo 3: 21 DOF, full-palm touch sensing. 🔵 Wuji Hand: 20 DOF, motors inside each finger. 🔵 1X NEO: 22 DOF, force sensing at every joint. Which grip wins?
English
46
173
946
141.6K
Gabb
Gabb@gabb0347·
@boxcardavid Every niche skillset has this issue. Just because this is not a niche you know, does not mean that domain specific terminology is not useful for the experts within a field. I for example could not explain to anyone what an offside means in football, i just hear people say it.
English
0
0
0
65
David Hansen 🇺🇸 🇳🇿
David Hansen 🇺🇸 🇳🇿@boxcardavid·
The worst part about robot hands is that they use human anatomy terms that I will never remember
Andrea@aesposito0

The thumb architecture is an interesting choice. 1X included thumb pronation/supination (axial rotation) but skipped abduction/adduction, which is the more common choice in humanoid hands. I loved how the @wuji_global hand solved this: it keeps ab/ad, then gets some axial rotation "for free" by angling the MCP/IP flexion axes relative to the CMC flexion axis.

English
3
2
94
7.3K
Gabb
Gabb@gabb0347·
@aesposito0 @BerntBornich In that specific video, maybe it does not show it that well. But look at the other videos, how would Neo's hand do opposition and pinches without CMC Y, and also be able to make the hand a flat paddle?
English
1
0
8
428
Owizzle
Owizzle@WillyB_303·
@gabb0347 @CosterDaniel @1x_tech Yes but by this and Bernt's logic with the House Select Committee on China, I can expect Håland to play for Team USA soccer in 2032 bc 1X is as American as 🍎🥧 🤣? Just teasing u.
English
1
0
2
48
1X
1X@1x_tech·
NEO’s Hands An API to the Physical World
English
1.2K
1.5K
16.1K
11.4M
Andrea
Andrea@aesposito0·
@gabb0347 @BerntBornich I’m talking about degrees of actuation: joints that can be moved independently. I count: - CMC axial rotation (pronation/supination) - CMC flexion/extension - (no 3rd CMC DOF) - MCP flexion/extension - IP flexion/extension Where’s the fifth?
Andrea tweet media
English
1
0
3
1.8K
Gabb
Gabb@gabb0347·
@aryind_ Opinion or fact? If fact, back it by data.
English
1
0
1
41
Gabb
Gabb@gabb0347·
@mikekalilmfg Interesting take. Let the future be the proof that is in the pudding. Tesla has claimed interesting things about their robot for the past year, i have not seen them making this much progress this fast.
English
1
0
2
138
Mike Kalil
Mike Kalil@mikekalilmfg·
1X’s NEO Hand Looks Incredible. The Economics Look Brutal. 1X is making huge claims about its robotic hand, calling it the “most advanced” in history. What they won’t say publicly, though, is that right now the economics do not add up. The Silicon Valley startup offered a closeup look at the end effector used by its humanoid robot NEO as it races to fulfill more than 10,000 preorders. The hand’s specs are objectively impressive on paper, and 1X shared footage to back up its claims. Its marketing video showed NEO doing a range of tasks requiring human-level precision like screwing in a lightbulb and handling a wine glass. 1X has announced about $137 million in funding to date. That sounds like a lot until you compare it to its biggest US competitor, Figure AI, which raised more than $1 billion in 2025. It also pales in comparison to several Chinese startups like RobotEra, Linkerbot, EngineAI, Galbot, Agibot, NOETIX Robotics, and LimX Dynamics that have already raised more just in 2026. And then there’s Unitree, which has been approved for a Shanghai IPO that could raise around $619 million. And this may explain why 1X is making such a big deal about the hand right now. It needs investors to believe it has solved one of the hardest problems in robotics: useful, humanlike manipulation. Because without a much larger funding round, the economics are hard to square. The Terrifying Math of 1X NEO's Hands mikekalil.com/blog/1x-neo-ha…
Mike Kalil tweet media
Mike Kalil@mikekalilmfg

There’s no way they’re not selling these robots at a considerable loss. Hands in this range go for at least $20,000 alone normally especially outside China. Most humanoids on the market are also $20,000+. Even a Unitree G1, you’re not gonna find one for below $30,000 in the US and that’s the base model. “But the R1 is $5,000.” That’s an RC toy, stop it. This is a huge gamble. The total loss could be well above $500 million. With all the R&D and infrastructure they need, this can easily eat up their entire war chest and then some. Their deployment strategy is also untested. They say they’re hoping for full autonomy by 2027 but no one knows what will actually happen. They won’t admit it publicly but there is a very high chance of failure. If they do pull it off, it would be huge though. Or maybe they’ll get nationalized so Trump can turn NEOs into humanoid soldiers.

English
6
2
40
9.3K
Gabb
Gabb@gabb0347·
@simonkalouche That is the old hand, which was designed a long time ago, things are different now.
English
0
0
1
493
Daniel Coster
Daniel Coster@CosterDaniel·
@1x_tech Norway is not only dominating the World Cup but also hands 👋🏻 🤯
English
1
0
3
2.3K
Gabb
Gabb@gabb0347·
Having worked on engineering different robotics systems for 9+ years, I am both tendon-drive and world model pilled. Because intentionally flexible and compliant systems tend to outperform over engineered, rigid and overtuned systems. Especially in unstructured environments.
English
0
0
3
589